INDEX

Explanations

phrases related to professions or positions

oai_token-act-pair · gpt-3.5-turbo

references to academic titles and positions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

SPONSORED

-0.72

rats

-0.69

 hinges

-0.67

\'

-0.65

?!

-0.63

↵Âł

-0.63

 rigged

-0.62

 idiots

-0.61

 attracts

-0.60

bucks

-0.60

POSITIVE LOGITS

 spokesman

0.87

 spokeswoman

0.87

 researcher

0.84

 Managing

0.84

 sociology

0.84

 spokesperson

0.83

 managing

0.81

 historian

0.81

 vice

0.80

 consultant

0.80

Activations Density 0.130%

phrases related to professions or positions

references to academic titles and positions

No Comments

No Known Activations