INDEX

Explanations

locations or directions

oai_token-act-pair · gpt-3.5-turbo

directional words and phrases indicating vertical positioning

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 Correction

-0.66

icer

-0.64

/>

-0.63

<<

-0.62

Compl

-0.62

 RELEASE

-0.62

;;;;;;;;;;;;

-0.61

Bye

-0.61

¼

-0.61

Paper

-0.60

POSITIVE LOGITS

stellar

0.74

field

0.73

ricular

0.71

 recess

0.70

ridor

0.69

itals

0.69

sighted

0.69

ngth

0.68

fetched

0.68

 periphery

0.67

Activations Density 0.108%

locations or directions

directional words and phrases indicating vertical positioning

No Comments

No Known Activations