上游文档原文（英文）

Capacity Controller

deepseek-tui includes an opt-in capacity-aware context controller. In the default V4 path it is disabled, because its active interventions can rewrite the live prompt and break prefix-cache affinity. Treat it as telemetry or an experimental guardrail unless capacity.enabled = true is set explicitly.

Policy Overview

Each checkpoint computes:

• H_hat (runtime pressure proxy)
• C_hat (model capacity prior)
• slack = C_hat - H_hat
• dynamic slack profile over last N=8 observations

Runtime Pressure Proxy (H_hat)

• action_complexity_bits = log2(1 + action_count_this_turn)
• tool_complexity_bits = log2(1 + tool_calls_recent_window)
• ref_complexity_bits = log2(1 + unique_reference_ids_recent_window)
• context_pressure_bits = 6.0 * context_used_ratio

Formula:

H_hat = 0.35*action_complexity_bits + 0.30*tool_complexity_bits + 0.20*ref_complexity_bits + 0.15*context_pressure_bits

Capacity Prior (C_hat)

Per-model priors:

• deepseek_v3_2_chat = 3.9
• deepseek_v3_2_reasoner = 4.1
• deepseek_v4_pro = 3.5
• deepseek_v4_flash = 4.2
• fallback 3.8 (used for other DeepSeek IDs, including future releases)

Failure Probability

Using rolling profile fields:

• final_slack
• min_slack
• violation_ratio
• slack_volatility
• slack_drop

Formula:

z = -1.65*final_slack -0.85*min_slack +1.35*violation_ratio +0.70*slack_volatility +0.28*slack_drop -0.12

p_fail = sigmoid(z) clamped to [0,1].

Risk bands:

• low: p_fail <= low_risk_max
• medium: p_fail <= medium_risk_max
• high: otherwise

Action mapping when the controller is explicitly enabled:

• low -> NoIntervention
• medium -> TargetedContextRefresh
• high + severe dynamics (min_slack <= severe_min_slack or violation_ratio >= severe_violation_ratio) -> VerifyAndReplan
• otherwise high -> VerifyWithToolReplay

Checkpoints

When enabled, the engine evaluates controller policy at:

1. Pre-request checkpoint (before MessageRequest assembly).
2. Post-tool checkpoint (after tool result append).
3. Error-escalation checkpoint (tool error streak path).

Interventions

Interventions are not part of the default v0.7.5 V4 path. The default path is: append messages, preserve prefix-cache reuse, suggest manual /compact near real model pressure, and use overflow recovery only if the request would exceed the model input budget.

TargetedContextRefresh

• Runs compaction (compact_messages_safe) when possible.
• Falls back to local trim if compaction path fails.
• Persists canonical state.
• Replaces long-tail active context with compact canonical prompt + memory pointer.

VerifyWithToolReplay

• Replays one read-only critical tool call from recent turn context.
• Appends verification note with pass/fail + diff summary.
• On replay conflict/error, marks escalation candidate and disables replay for current turn.

VerifyAndReplan

• Persists canonical snapshot.
• Clears volatile prompt tail while preserving latest user ask and latest verification note.
• Injects canonical replan instruction into system prompt.
• Continues turn loop from compact canonical state.

Safety Controls

• Max one intervention per turn.
• Cooldowns for refresh and replan.
• Replay budget per turn (max_replay_per_turn).
• Fail-open behavior when controller inputs are unavailable.
• Compaction/replay failures are logged; turn continues.

Memory Store

Path:

• DEEPSEEK_CAPACITY_MEMORY_DIR (if set)
• otherwise ~/.deepseek/memory/<session_id>.jsonl
• fallback: <workspace>/.deepseek/memory/<session_id>.jsonl when home path is unavailable/unwritable

Record fields:

• id, ts, turn_index, action_trigger
• h_hat, c_hat, slack, risk_band
• canonical_state
• source_message_ids
• optional replay_info

Loader utility supports fetching last K snapshots for rehydration.

Configuration

[capacity] keys:

• enabled (default false)
• low_risk_max (default 0.50)
• medium_risk_max (default 0.62)
• severe_min_slack (default -0.25)
• severe_violation_ratio (default 0.40)
• refresh_cooldown_turns (default 6)
• replan_cooldown_turns (default 5)
• max_replay_per_turn (default 1)
• min_turns_before_guardrail (default 4)
• profile_window (default 8)
• deepseek_v3_2_chat_prior (default 3.9)
• deepseek_v3_2_reasoner_prior (default 4.1)
• deepseek_v4_pro_prior (default 3.5)
• deepseek_v4_flash_prior (default 4.2)
• fallback_default_prior (default 3.8)

Equivalent environment overrides are available with DEEPSEEK_CAPACITY_*.

回到顶部