INDEX

Explanations

references to specific locations or names, particularly "Holland"

oai_token-act-pair · gpt-3.5-turbo

mentions of geographic locations, particularly "Holland."

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

ctic

-0.84

culation

-0.84

ramid

-0.82

vae

-0.82

ccording

-0.81

kered

-0.77

ãĥ£

-0.76

nces

-0.76

ormal

-0.74

arettes

-0.74

POSITIVE LOGITS

 Holland

1.21

stadt

0.83

ijn

0.82

 Gaal

0.80

inson

0.79

 Jarvis

0.77

cott

0.77

 Gardens

0.76

 Spit

0.76

shire

0.75

Activations Density 0.008%

references to specific locations or names, particularly "Holland"

mentions of geographic locations, particularly "Holland."

No Comments

No Known Activations