INDEX

Explanations

words related to dispelling something or breaking into pieces

oai_token-act-pair · gpt-3.5-turbo

terms related to dispelling misconceptions and experiences of destruction or damage

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

gee

-0.90

ecake

-0.84

essee

-0.83

gage

-0.74

 nomine

-0.71

iquid

-0.71

tarian

-0.69

asus

-0.67

ung

-0.66

 compose

-0.65

POSITIVE LOGITS

URA

0.86

ãĤ©

0.85

arov

0.80

flies

0.77

 shattered

0.74

tons

0.73

actionDate

0.72

Scroll

0.71

 illusions

0.70

ECK

0.70

Activations Density 0.038%

words related to dispelling something or breaking into pieces

terms related to dispelling misconceptions and experiences of destruction or damage

No Comments

No Known Activations