INDEX

Explanations

phrases related to a broad or wide range of topics or issues

oai_token-act-pair · gpt-3.5-turbo

phrases indicating extensive scope or range in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 machine

-0.78

 stocks

-0.71

 capital

-0.70

 capitals

-0.67

 static

-0.67

 blob

-0.66

ut

-0.64

sto

-0.64

 crest

-0.64

 default

-0.64

POSITIVE LOGITS

ranging

4.16

reaching

1.93

ranged

1.83

rang

1.24

range

1.15

igning

0.99

seeing

0.97

dating

0.97

talking

0.96

joining

0.93

Activations Density 0.019%

phrases related to a broad or wide range of topics or issues

phrases indicating extensive scope or range in various contexts

No Comments

No Known Activations