INDEX

Explanations

images or symbols associated with guidance or prominence

oai_token-act-pair · gpt-3.5-turbo

references to significant landmarks or symbols

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

ivan

-0.83

ean

-0.79

ordan

-0.71

andro

-0.69

dos

-0.69

ulu

-0.68

EVA

-0.66

cube

-0.66

RAW

-0.64

ktop

-0.62

POSITIVE LOGITS

 beacon

1.18

 Beacon

1.07

bom

0.82

 transmitter

0.77

-+-+

0.77

 boosters

0.76

ãĤ¤ãĥĪ

0.74

Cyn

0.71

ically

0.70

 outpost

0.70

Activations Density 0.008%

images or symbols associated with guidance or prominence

references to significant landmarks or symbols

No Comments

No Known Activations