INDEX

Explanations

terms related to spatial directions, specifically 'left' and 'right'

oai_token-act-pair · gpt-3.5-turbo

references to directional terms and their relationships

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ĸļ

-0.82

aton

-0.72

odium

-0.71

onite

-0.70

acus

-0.70

gat

-0.69

ynt

-0.69

orious

-0.68

wana

-0.67

exempt

-0.67

POSITIVE LOGITS

 sided

0.96

 hemisphere

0.92

 wing

0.92

 flank

0.91

 aligned

0.83

 lateral

0.82

 aisle

0.82

 side

0.81

 handed

0.80

 triangles

0.79

Activations Density 0.069%

terms related to spatial directions, specifically 'left' and 'right'

references to directional terms and their relationships

No Comments

No Known Activations