INDEX

Explanations

references to a specific location, Fort Detrick, within various contexts

oai_token-act-pair · gpt-3.5-turbo

references to Detroit and its related entities

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

manship

-0.89

xual

-0.88

nown

-0.79

ammy

-0.76

hetti

-0.74

externalActionCode

-0.69

lihood

-0.66

#$

-0.65

heit

-0.62

 deaf

-0.62

POSITIVE LOGITS

ector

1.18

ention

1.12

ected

1.12

rans

1.11

roit

1.06

ailed

1.03

ective

1.02

ainers

1.01

ainer

0.98

ection

0.97

Activations Density 0.014%

references to a specific location, Fort Detrick, within various contexts

references to Detroit and its related entities

No Comments

No Known Activations