INDEX

Explanations

numerical data presented in a list format

oai_token-act-pair · gpt-3.5-turbo

commas in numerical data presentations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 newsp

-0.71

 likeness

-0.68

 enthusi

-0.66

 neighb

-0.62

 nomine

-0.60

 advoc

-0.60

 fundament

-0.60

 disadvant

-0.59

 existence

-0.58

 philos

-0.58

POSITIVE LOGITS

1.57

1.55

1.52

1.48

1.47

1.45

Activations Density 0.089%

numerical data presented in a list format

commas in numerical data presentations

No Comments

No Known Activations