INDEX

Explanations

technical terms related to software development and system testing

oai_token-act-pair · gpt-3.5-turbo

elements related to structured data or navigation

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

amins

-0.87

eland

-0.73

oche

-0.73

reb

-0.73

 Galile

-0.72

rette

-0.72

quez

-0.68

elly

-0.68

lys

-0.68

jured

-0.66

POSITIVE LOGITS

 ---------

0.86

 -------

0.81

=~=~

0.80

-->

0.75

 <!--

0.73

ADE

0.71

/-

0.71

BACK

0.69

-->

0.68

 ----------------

0.68

Activations Density 0.016%

technical terms related to software development and system testing

elements related to structured data or navigation

No Comments

No Known Activations