INDEX

Explanations

words related to governmental entities and programs

oai_token-act-pair · gpt-3.5-turbo

topics related to government and politics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Deity

-0.83

pload

-0.74

Sym

-0.71

 Schl

-0.70

ragon

-0.70

ï¸

-0.68

 Pirate

-0.67

 Candle

-0.66

Wave

-0.66

 Stones

-0.66

POSITIVE LOGITS

 appoint

0.88

 appointed

0.80

 subsid

0.78

 decree

0.74

 funded

0.72

 mend

0.71

kefeller

0.70

ĳå£«

0.68

 oversight

0.67

 authorised

0.67

Activations Density 0.419%

words related to governmental entities and programs

topics related to government and politics

No Comments

No Known Activations