INDEX

Explanations

places or structures where things are located

oai_token-act-pair · gpt-3.5-turbo

instances of the word "housed."

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

 Carney

-0.66

vious

-0.65

uyomi

-0.64

reme

-0.63

xxx

-0.61

isy

-0.61

alias

-0.60

reshold

-0.59

 Snow

-0.58

eds

-0.58

POSITIVE LOGITS

keeping

0.85

wives

0.80

Īè

0.79

maid

0.79

owship

0.78

byss

0.76

hips

0.75

chool

0.74

ousing

0.74

ãĥķãĤ¡

0.73

Activations Density 0.025%

places or structures where things are located

instances of the word "housed."

No Comments

No Known Activations