INDEX

Explanations

phrases related to covert operations or undercover activities

oai_token-act-pair · gpt-3.5-turbo

references to undercover operations and agents

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ettle

-0.82

pite

-0.70

reciation

-0.70

itsch

-0.70

atomic

-0.69

estone

-0.69

plex

-0.69

yards

-0.69

ipples

-0.68

ð

-0.68

POSITIVE LOGITS

 informant

1.03

 informants

0.94

 operatives

0.93

 undercover

0.92

 infiltr

0.87

 sting

0.86

FBI

0.84

 investigator

0.79

 infiltration

0.78

CIA

0.75

Activations Density 0.075%

phrases related to covert operations or undercover activities

references to undercover operations and agents

No Comments

No Known Activations