INDEX

Explanations

phrases related to some items being out of a place or context

oai_token-act-pair · gpt-3.5-turbo

instances of the word "out."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 arsen

-0.67

Pry

-0.64

 grooming

-0.63

 refined

-0.63

 bonded

-0.62

 melting

-0.61

age

-0.61

 trem

-0.61

 Puzzles

-0.58

 irrad

-0.58

POSITIVE LOGITS

lier

1.06

door

1.04

doors

1.02

casts

1.00

stretched

0.93

out

0.93

outs

0.93

dated

0.92

lander

0.90

fitted

0.89

Activations Density 0.014%

phrases related to some items being out of a place or context

instances of the word "out."

No Comments

No Known Activations