INDEX

Explanations

adjectives related to textures

oai_token-act-pair · gpt-3.5-turbo

words with the suffix '-y' or '-ed'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

employ

-0.75

CVE

-0.70

REP

-0.69

esson

-0.69

 Leilan

-0.68

MRI

-0.67

SEA

-0.66

podcast

-0.66

igun

-0.66

GV

-0.63

POSITIVE LOGITS

 equivalents

0.78

 greens

0.75

 substr

0.75

 plains

0.74

 beverage

0.74

 slopes

0.74

 tones

0.73

 compounds

0.73

 waters

0.71

 beverages

0.70

Activations Density 0.116%

adjectives related to textures

words with the suffix '-y' or '-ed'

No Comments

No Known Activations