INDEX

Explanations

locations or scenes described with the word "pan" (short for panoramic) followed by a number

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "pan" and related terms

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

IELD

-0.78

 Annotations

-0.73

IGH

-0.64

æ©Ł

-0.64

IFE

-0.63

INGTON

-0.62

ITED

-0.60

 Norn

-0.59

ORD

-0.59

URI

-0.58

POSITIVE LOGITS

orama

1.13

handler

1.07

acea

1.05

eful

0.99

efully

0.93

ter

0.93

artisan

0.90

icles

0.90

shots

0.89

ocular

0.88

Activations Density 0.013%

locations or scenes described with the word "pan" (short for panoramic) followed by a number

occurrences of the word "pan" and related terms

No Comments

No Known Activations