INDEX

Explanations

names and terms related to specific individuals

oai_token-act-pair · gpt-3.5-turbo

names frequently associated with significant individuals and leaders

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

tics

-0.89

icum

-0.81

erie

-0.79

ographic

-0.75

chini

-0.74

furt

-0.73

ication

-0.72

onyms

-0.72

ographers

-0.71

ures

-0.70

POSITIVE LOGITS

vernment

0.73

 incorpor

0.73

rane

0.71

itle

0.67

 dred

0.66

++++++++++++++++

0.66

enegger

0.66

vre

0.65

inness

0.65

 Mahar

0.63

Activations Density 0.115%

names and terms related to specific individuals

names frequently associated with significant individuals and leaders

No Comments

No Known Activations