INDEX

Explanations

unique identification numbers or cards

oai_token-act-pair · gpt-3.5-turbo

terms related to identification documents and IDs

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

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

unin

-0.79

iful

-0.73

sei

-0.71

cffff

-0.68

theless

-0.68

ovie

-0.66

Bed

-0.65

 Silence

-0.64

untled

-0.64

storms

-0.63

POSITIVE LOGITS

 card

1.02

entity

1.01

 cards

1.01

iots

1.01

ENT

1.00

aho

0.96

 badges

0.95

 badge

0.94

EA

0.91

ocument

0.89

Activations Density 0.032%

unique identification numbers or cards

terms related to identification documents and IDs

No Comments

No Known Activations