INDEX

Explanations

expressions of involvement or participation

oai_token-act-pair · gpt-3.5-turbo

phrases indicating participation or inclusion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 decay

-0.68

 dislike

-0.64

 expire

-0.63

 botched

-0.62

usterity

-0.61

 Purg

-0.60

cheat

-0.60

 Cooldown

-0.60

abuse

-0.59

 mell

-0.58

POSITIVE LOGITS

 part

1.62

part

1.15

 integral

1.05

 contributors

1.01

Part

1.00

 involved

0.99

Reviewer

0.97

 contributing

0.97

 instrumental

0.96

 participant

0.96

Activations Density 0.371%

expressions of involvement or participation

phrases indicating participation or inclusion

No Comments

No Known Activations