INDEX

Explanations

keywords related to weight, such as measurements or weight-related terms

oai_token-act-pair · gpt-3.5-turbo

references to weight and physical measurements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 paras

-0.71

 emer

-0.70

 harassing

-0.70

 harass

-0.69

 agent

-0.69

 pear

-0.69

 fault

-0.68

 influ

-0.68

 precip

-0.67

 sequ

-0.67

POSITIVE LOGITS

 Weight

3.00

Weight

2.74

 Height

1.36

roud

1.29

 Suff

1.15

weight

1.15

 Dimensions

1.08

Height

1.08

Rating

1.06

 Kens

1.03

Activations Density 0.031%

keywords related to weight, such as measurements or weight-related terms

references to weight and physical measurements

No Comments

No Known Activations