INDEX

Explanations

the word "fans"

oai_token-act-pair · gpt-3.5-turbo

references to fans and their importance or involvement

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

nom

-0.70

 srfAttach

-0.66

lished

-0.65

olon

-0.64

ateral

-0.62

ãģĨ

-0.62

Prosecut

-0.62

 Surgery

-0.62

EDIT

-0.61

vet

-0.61

POSITIVE LOGITS

haw

0.93

atics

0.93

atical

0.88

ubs

0.87

boo

0.85

 cheering

0.84

 Atmosp

0.84

verson

0.83

atti

0.82

boys

0.79

Activations Density 0.030%

the word "fans"

references to fans and their importance or involvement

No Comments

No Known Activations