INDEX

Explanations

specific numerical patterns

oai_token-act-pair · gpt-3.5-turbo

numeric codes or identifiers related to various entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

achus

-0.80

BOOK

-0.76

skirts

-0.73

uler

-0.72

prises

-0.71

 compr

-0.70

oÄŁ

-0.67

hend

-0.67

uran

-0.66

mble

-0.66

POSITIVE LOGITS

 ILCS

1.08

0.88

lav

0.85

«ĺ

0.83

MHz

0.82

0.81

0.73

wagen

0.72

Activations Density 0.049%

specific numerical patterns

numeric codes or identifiers related to various entities

No Comments

No Known Activations