INDEX

Explanations

phrases or words related to secrets

oai_token-act-pair · gpt-3.5-turbo

references to "secret" or "classified" information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

gaard

-0.84

merce

-0.78

annis

-0.77

ð

-0.74

ulf

-0.72

odcast

-0.70

enburg

-0.68

kins

-0.66

adjusted

-0.65

boa

-0.64

POSITIVE LOGITS

ariat

1.47

arial

1.43

 handshake

0.95

 sauce

0.92

 stash

0.87

 ingredient

0.84

rets

0.78

iveness

0.77

 cabal

0.76

ively

0.75

Activations Density 0.042%

phrases or words related to secrets

references to "secret" or "classified" information

No Comments

No Known Activations