INDEX

Explanations

names or titles of individuals associated with authority or leadership positions

oai_token-act-pair · gpt-3.5-turbo

titles and ranks of authoritative figures

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 compr

-0.75

*/(

-0.74

icult

-0.73

 fracture

-0.71

aukee

-0.71

 overshadow

-0.70

 destro

-0.69

sole

-0.69

crew

-0.69

 reper

-0.68

POSITIVE LOGITS

 Richard

1.18

 Joseph

1.13

 David

1.12

 William

1.10

 Andrew

1.10

 Reed

1.08

 Gordon

1.08

 Theodore

1.07

 George

1.07

 Chomsky

1.07

Activations Density 0.147%

names or titles of individuals associated with authority or leadership positions

titles and ranks of authoritative figures

No Comments

No Known Activations