INDEX

Explanations

positive expressions or compliments

oai_token-act-pair · gpt-3.5-turbo

positive qualitative evaluations or experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

idden

-0.78

":[

-0.77

oti

-0.76

Newsletter

-0.74

conservancy

-0.72

artney

-0.71

owship

-0.71

vertisement

-0.68

ographers

-0.68

obook

-0.68

POSITIVE LOGITS

huh

1.16

 congr

1.03

eh

0.96

 dude

0.89

 coincidence

0.85

 lucky

0.81

 kidding

0.80

 gotta

0.79

 typo

0.79

tho

0.79

Activations Density 0.659%

positive expressions or compliments

positive qualitative evaluations or experiences

No Comments

No Known Activations