INDEX

Explanations

geographical locations, particularly highlighting places ending in "om"

oai_token-act-pair · gpt-3.5-turbo

proper nouns related to people and place names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 reversible

-0.68

BILITIES

-0.67

 conflic

-0.67

 suites

-0.66

nect

-0.64

 rapport

-0.64

 pestic

-0.63

 misdem

-0.63

 href

-0.62

 Relationship

-0.61

POSITIVE LOGITS

warts

1.02

lich

0.89

gow

0.75

schild

0.72

ratulations

0.70

she

0.69

adesh

0.68

 Saban

0.67

ooked

0.67

idan

0.65

Activations Density 0.205%

geographical locations, particularly highlighting places ending in "om"

proper nouns related to people and place names

No Comments

No Known Activations