INDEX

Explanations

words related to weight or weight management

oai_token-act-pair · gpt-3.5-turbo

references to weight and related metrics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ICLE

-0.77

BLE

-0.71

xon

-0.70

RPG

-0.69

Suc

-0.69

etheless

-0.68

VIS

-0.68

cffffcc

-0.67

 Noir

-0.66

 Gutenberg

-0.66

POSITIVE LOGITS

lifting

1.27

lif

0.98

 lifting

0.95

weight

0.94

lessness

0.91

weights

0.89

 weights

0.87

loss

0.84

cake

0.83

 weight

0.83

Activations Density 0.032%

words related to weight or weight management

references to weight and related metrics

No Comments

No Known Activations