INDEX

Explanations

locations or organizations, perhaps related to a specific region

oai_token-act-pair · gpt-3.5-turbo

references to geographical locations, particularly related to "rica" and "diocese."

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

nut

-0.95

ington

-0.85

ton

-0.83

ty

-0.81

bart

-0.80

holes

-0.79

nuts

-0.77

ties

-0.77

sticks

-0.76

geon

-0.74

POSITIVE LOGITS

ANA

1.13

 Albania

0.96

ADA

0.93

yss

0.89

ocese

0.88

 Rwanda

0.81

 Alger

0.80

AAF

0.74

imbabwe

0.74

 Scotia

0.74

Activations Density 0.030%

locations or organizations, perhaps related to a specific region

references to geographical locations, particularly related to "rica" and "diocese."

No Comments

No Known Activations