INDEX

Explanations

instances of the word "feedback" in the text

oai_token-act-pair · gpt-3.5-turbo

references to user or player feedback

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

neys

-0.74

wikipedia

-0.73

chin

-0.72

amina

-0.71

annis

-0.70

ffe

-0.66

eni

-0.66

adem

-0.66

Jew

-0.64

asar

-0.63

POSITIVE LOGITS

 loops

1.07

 feedback

1.06

 Feedback

1.04

 loop

0.89

 assurance

0.77

 overload

0.75

ible

0.75

 booths

0.72

 testers

0.72

urai

0.70

Activations Density 0.062%

instances of the word "feedback" in the text

references to user or player feedback

No Comments

No Known Activations