INDEX

Explanations

patterns or structured arrangements in text

oai_token-act-pair · gpt-3.5-turbo

references to recurring themes or trends

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ascular

-0.73

avez

-0.71

rican

-0.70

zona

-0.70

vez

-0.70

omez

-0.70

ossip

-0.70

IVERS

-0.68

ayson

-0.67

hiro

-0.66

POSITIVE LOGITS

 pattern

0.99

ĸļ

0.99

 patterns

0.97

 Pattern

0.93

pattern

0.91

 Patterns

0.90

Pattern

0.89

eering

0.86

atile

0.83

gradient

0.83

Activations Density 0.014%

patterns or structured arrangements in text

references to recurring themes or trends

No Comments

No Known Activations