INDEX

Explanations

phrases referring to specific objects or concepts that are important or significant to the person or group mentioned in the text

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing singular importance

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

ilings

-0.72

ashington

-0.69

utics

-0.68

ategor

-0.67

dinand

-0.66

ciples

-0.65

ornings

-0.65

zhen

-0.64

odus

-0.64

otypes

-0.64

POSITIVE LOGITS

 conceivable

0.93

 imaginable

0.82

 elig

0.81

able

0.81

 available

0.79

 redeem

0.79

INO

0.78

 ever

0.77

 feasible

0.76

 authorized

0.75

Activations Density 0.121%

phrases referring to specific objects or concepts that are important or significant to the person or group mentioned in the text

phrases emphasizing singular importance

No Comments

No Known Activations