INDEX

Explanations

activities related to exerting effort or pressure

oai_token-act-pair · gpt-3.5-turbo

variations of the word "push."

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

abol

-0.72

Interstitial

-0.70

 Recogn

-0.70

omial

-0.68

 Receiver

-0.66

uster

-0.65

abad

-0.64

enfranch

-0.59

 Seym

-0.59

 Chaff

-0.58

POSITIVE LOGITS

chairs

0.91

 pushing

0.85

 push

0.85

 pushed

0.82

back

0.82

push

0.78

 harder

0.77

boxes

0.77

 toward

0.77

 forward

0.76

Activations Density 0.031%

activities related to exerting effort or pressure

variations of the word "push."

No Comments

No Known Activations