INDEX

Explanations

adjectives related to texture or consistency

oai_token-act-pair · gpt-3.5-turbo

words ending with the suffix "y."

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

ajor

-0.88

 Examiner

-0.82

itutional

-0.73

rongh

-0.71

raltar

-0.69

bernatorial

-0.69

igslist

-0.66

rador

-0.65

arians

-0.64

erver

-0.62

POSITIVE LOGITS

ielding

0.92

ield

0.87

ness

0.86

bean

0.82

 eyed

0.80

lly

0.77

bye

0.74

oes

0.73

balls

0.71

 sounding

0.70

Activations Density 0.041%

adjectives related to texture or consistency

words ending with the suffix "y."

No Comments

No Known Activations