INDEX

Explanations

instances of the word "ale"

oai_token-act-pair · gpt-3.5-turbo

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

nesses

-0.83

ness

-0.82

nl

-0.77

staff

-0.71

NESS

-0.68

ENTION

-0.65

 Beir

-0.64

sit

-0.62

ulates

-0.62

olicy

-0.61

POSITIVE LOGITS

cki

1.24

xit

1.20

ppo

1.07

uca

1.01

Ã±o

0.92

ttes

0.91

esi

0.91

ea

0.91

lla

0.83

zzi

0.83

Activations Density 0.050%

instances of the word "ale"

No Comments

No Known Activations