INDEX

Explanations

names or references related to secret agents

oai_token-act-pair · gpt-3.5-turbo

mentions of agents, particularly in law enforcement or espionage contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

etooth

-0.79

FACE

-0.76

issance

-0.73

ership

-0.72

lihood

-0.72

jri

-0.70

Ł

-0.67

frey

-0.66

rise

-0.65

orld

-0.64

POSITIVE LOGITS

 prov

1.12

agent

0.90

 agent

0.90

 Agents

0.88

 posing

0.88

agents

0.87

 agents

0.86

 Agent

0.86

sov

0.80

Agent

0.77

Activations Density 0.049%

names or references related to secret agents

mentions of agents, particularly in law enforcement or espionage contexts

No Comments

No Known Activations