INDEX

Explanations

words related to specific organizations, particularly associations

oai_token-act-pair · gpt-3.5-turbo

references to various associations or organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

gone

-0.82

agram

-0.71

tro

-0.71

things

-0.70

lasses

-0.68

ãĤĮ

-0.67

=-=-=-=-=-=-=-=-

-0.64

ths

-0.64

meyer

-0.61

oÄŁ

-0.61

POSITIVE LOGITS

 Association

0.92

UTH

0.84

 Confederation

0.82

eer

0.80

SPA

0.79

 Agreement

0.79

uthor

0.77

ociation

0.76

 Associ

0.76

 Board

0.76

Activations Density 0.021%

words related to specific organizations, particularly associations

references to various associations or organizations

No Comments

No Known Activations