INDEX

Explanations

names and titles

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals or entities, particularly those related to "Vic" and "Tru."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Kingdom

-0.76

 STATES

-0.67

 Devil

-0.63

 States

-0.62

 recruitment

-0.62

 Erie

-0.62

 Arab

-0.62

 Nations

-0.61

 Lakes

-0.59

 refugee

-0.59

POSITIVE LOGITS

sense

0.86

iple

0.80

plin

0.79

cipled

0.79

swer

0.77

sa

0.76

itto

0.76

ILLE

0.75

forward

0.74

iren

0.74

Activations Density 0.090%

names and titles

references to specific individuals or entities, particularly those related to "Vic" and "Tru."

No Comments

No Known Activations