INDEX

Explanations

names of locations or organizations with the abbreviation "MT"

oai_token-act-pair · gpt-3.5-turbo

mention of media technology or media-related terms and entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

hold

-0.88

 Horowitz

-0.72

amus

-0.71

pta

-0.68

holes

-0.64

 Hitman

-0.64

 Fulton

-0.64

Boe

-0.63

 Spencer

-0.62

Pog

-0.61

POSITIVE LOGITS

MT

1.06

ypes

0.85

orsi

0.84

ullah

0.81

ument

0.80

olkien

0.80

ormon

0.77

TL

0.77

ertodd

0.77

TON

0.77

Activations Density 0.007%

names of locations or organizations with the abbreviation "MT"

mention of media technology or media-related terms and entities

No Comments

No Known Activations