INDEX

Explanations

binary choices or decisions in text

oai_token-act-pair · gpt-3.5-turbo

negative responses or indicators of disapproval

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 anim

-0.64

 transpl

-0.62

bes

-0.62

 wand

-0.60

 arts

-0.60

 Middle

-0.58

 Wand

-0.58

 appet

-0.57

 handcuffs

-0.57

effic

-0.57

POSITIVE LOGITS

NO

4.13

NO

2.49

YES

1.75

no

1.71

YES

1.44

NI

1.30

ANY

1.20

VO

1.19

SO

1.17

DON

1.17

Activations Density 0.017%

binary choices or decisions in text

negative responses or indicators of disapproval

No Comments

No Known Activations