INDEX

Explanations

names of geographical locations, specifically focusing on a specific location that contains the text "gon"

oai_token-act-pair · gpt-3.5-turbo

references to a specific geographic location or area

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 credibility

-0.76

 replay

-0.69

 Marriott

-0.67

RF

-0.65

 cred

-0.65

 lift

-0.64

 shred

-0.63

 Establishment

-0.62

Ref

-0.62

Ans

-0.61

POSITIVE LOGITS

gon

4.72

gor

1.26

gon

1.19

gas

1.05

gian

1.03

gam

1.00

nor

0.99

gom

0.98

gyn

0.98

morph

0.97

Activations Density 0.021%

names of geographical locations, specifically focusing on a specific location that contains the text "gon"

references to a specific geographic location or area

No Comments

No Known Activations