INDEX

Explanations

locations described as intersections

oai_token-act-pair · gpt-3.5-turbo

references to various intersections, especially in a physical or metaphorical context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

sonian

-0.80

wagen

-0.79

apons

-0.79

warm

-0.73

unker

-0.71

IER

-0.68

lee

-0.68

yre

-0.67

iken

-0.66

pload

-0.66

POSITIVE LOGITS

 intersection

0.87

 intersections

0.84

 Sakuya

0.76

 Beir

0.72

reon

0.71

aminer

0.69

lane

0.69

 Ceres

0.68

 Frie

0.68

 between

0.68

Activations Density 0.015%

locations described as intersections

references to various intersections, especially in a physical or metaphorical context

No Comments

No Known Activations