INDEX

Explanations

phrases related to specific movies or TV shows

oai_token-act-pair · gpt-3.5-turbo

titles or names associated with animated series and films

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

 themselves

-0.83

tein

-0.72

 himself

-0.71

 Himself

-0.65

 yourselves

-0.64

idates

-0.64

 Canaver

-0.62

hire

-0.62

atum

-0.61

arnaev

-0.61

POSITIVE LOGITS

 colleague

0.86

elist

0.84

 friends

0.77

 myself

0.77

favorite

0.76

Favorite

0.76

husband

0.76

 colleagues

0.73

 inbox

0.72

mates

0.71

Activations Density 0.274%

phrases related to specific movies or TV shows

titles or names associated with animated series and films

No Comments

No Known Activations