INDEX

Explanations

words related to lineups or series of items

oai_token-act-pair · gpt-3.5-turbo

references to various types of lineups or groupings

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

nai

-0.92

birth

-0.83

imb

-0.77

rir

-0.75

bara

-0.73

ryn

-0.72

mit

-0.68

cknowled

-0.68

hig

-0.67

versely

-0.67

POSITIVE LOGITS

 lineup

0.96

XI

0.89

selection

0.81

aurus

0.78

XV

0.76

 favourites

0.71

 swaps

0.70

 mates

0.66

 choices

0.66

 albums

0.63

Activations Density 0.016%

words related to lineups or series of items

references to various types of lineups or groupings

No Comments

No Known Activations