INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ĪĴ

-0.76

ctors

-0.75

istrate

-0.73

ournal

-0.73

ajo

-0.70

onse

-0.70

İĭ

-0.69

ngth

-0.69

annel

-0.68

ometime

-0.68

POSITIVE LOGITS

hab

0.78

isms

0.74

 Zoro

0.70

ised

0.66

rol

0.64

utter

0.64

mentioned

0.63

 INST

0.63

Philipp

0.61

tra

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.