INDEX

Explanations

questions asking for information or clarification

oai_token-act-pair · gpt-3.5-turbo

phrases or questions specifically related to location

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

 hopes

-0.65

 soon

-0.58

 fabric

-0.57

illon

-0.57

 fiber

-0.57

 blockers

-0.56

ima

-0.56

 rights

-0.56

 circulation

-0.55

 medium

-0.54

POSITIVE LOGITS

 Where

3.16

Where

2.35

where

1.95

 WHERE

1.89

WHERE

1.57

 where

1.46

 Somewhere

1.36

Why

1.33

Who

1.29

 What

1.28

Activations Density 0.012%

questions asking for information or clarification

phrases or questions specifically related to location

No Comments

No Known Activations