INDEX

Explanations

mentions of the word "Out" followed by a number, possibly related to locations, titles, or names

oai_token-act-pair · gpt-3.5-turbo

references to the term "Out" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 arsen

-0.75

iosity

-0.70

avorite

-0.70

 compr

-0.69

-0.66

 trem

-0.66

uzzle

-0.65

BILITY

-0.65

EY

-0.64

ä»£

-0.63

POSITIVE LOGITS

stretched

0.98

landish

0.97

rage

0.96

raged

0.95

breaks

0.94

doors

0.94

skirts

0.94

numbered

0.94

reach

0.92

casts

0.92

Activations Density 0.019%

mentions of the word "Out" followed by a number, possibly related to locations, titles, or names

references to the term "Out" in various contexts

No Comments

No Known Activations