INDEX

Explanations

the concept of comparison, particularly focusing on the word "lesser."

oai_token-act-pair · gpt-3.5-turbo

references to lesser-known or undervalued entities or individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

VR

-0.73

DK

-0.73

MM

-0.71

 Lobby

-0.68

Gro

-0.67

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

-0.67

alys

-0.66

 Pinball

-0.66

Closure

-0.66

 Population

-0.65

POSITIVE LOGITS

etheless

0.97

 than

0.95

 lesser

0.94

nesota

0.86

 minded

0.85

 sized

0.82

than

0.79

parts

0.70

nces

0.70

 "$:/

0.69

Activations Density 0.006%

the concept of comparison, particularly focusing on the word "lesser."

references to lesser-known or undervalued entities or individuals

No Comments

No Known Activations