INDEX

Explanations

references to geographical locations, specifically states

oai_token-act-pair · gpt-3.5-turbo

mentions of state-related concepts or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

poons

-0.79

icago

-0.78

eger

-0.77

ogle

-0.73

aurus

-0.73

ãĤ¨ãĥ«

-0.73

schild

-0.71

edin

-0.69

 Aires

-0.68

imon

-0.66

POSITIVE LOGITS

 legislature

1.44

 Legislature

1.32

 legislatures

1.23

 senate

1.11

 legislator

1.07

wide

1.05

 constitution

1.00

 senator

0.97

 trooper

0.95

 legislators

0.93

Activations Density 0.061%

references to geographical locations, specifically states

mentions of state-related concepts or entities

No Comments

No Known Activations