INDEX

Explanations

phrases about the effectiveness or impact of actions or decisions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "effectively" indicating actions or outcomes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 enthus

-0.73

Ey

-0.69

 Craw

-0.66

andro

-0.66

 favourites

-0.65

erer

-0.64

AMY

-0.64

 Fever

-0.62

erers

-0.61

TERN

-0.60

POSITIVE LOGITS

Ĥİ

0.90

 eliminate

0.73

 unemploy

0.71

 communicate

0.70

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

0.70

initions

0.68

 bankrupt

0.67

merce

0.67

lyak

0.66

 silenced

0.65

Activations Density 0.022%

phrases about the effectiveness or impact of actions or decisions

instances of the word "effectively" indicating actions or outcomes

No Comments

No Known Activations