INDEX

Explanations

questions

oai_token-act-pair · gpt-3.5-turbo

question and answer formats in the text

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

 outweigh

-0.70

atten

-0.68

 outwe

-0.68

guards

-0.68

ords

-0.66

 coales

-0.66

 comple

-0.64

oval

-0.63

 bloom

-0.61

umblr

-0.60

POSITIVE LOGITS

Why

0.97

WHY

0.94

 What

0.92

 Explain

0.92

Why

0.87

Hi

0.83

 Hello

0.83

Hello

0.80

How

0.79

 Suppose

0.79

Activations Density 0.054%

questions

question and answer formats in the text

No Comments

No Known Activations