INDEX

Explanations

phrases expressing a contrast or negation

oai_token-act-pair · gpt-3.5-turbo

phrases indicating decreasing quantity or comparison

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

hill

-0.67

 hurd

-0.66

 dexter

-0.66

cox

-0.65

encies

-0.63

Run

-0.63

arthed

-0.61

aston

-0.61

iencies

-0.60

ricks

-0.60

POSITIVE LOGITS

anship

0.84

tainment

0.73

cially

0.71

 Solitaire

0.70

ports

0.66

 Levant

0.64

atically

0.64

 Locations

0.61

 Digest

0.61

 Ø§ÙĦ

0.60

Activations Density 0.360%

phrases expressing a contrast or negation

phrases indicating decreasing quantity or comparison

No Comments

No Known Activations