INDEX

Explanations

topics related to government, politics, or military intelligence

oai_token-act-pair · gpt-3.5-turbo

details regarding government or legal documents and significant program information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

erential

-0.61

etitive

-0.56

 cautiously

-0.55

spective

-0.55

spect

-0.54

astical

-0.53

xtap

-0.53

earcher

-0.53

 prompted

-0.52

imb

-0.52

POSITIVE LOGITS

 goddamn

0.91

 fucking

0.87

 fuckin

0.86

 shit

0.85

 crap

0.85

 ..."

0.81

 gonna

0.80

 shitty

0.80

 FUCK

0.79

 damn

0.79

Activations Density 1.732%

topics related to government, politics, or military intelligence

details regarding government or legal documents and significant program information

No Comments

No Known Activations