INDEX

Explanations

pairs or groups of entities or topics discussed together

oai_token-act-pair · gpt-3.5-turbo

references to duality or items expressed in pairs

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

renheit

-0.84

ushima

-0.75

uable

-0.75

ugu

-0.75

frac

-0.72

Charg

-0.70

naire

-0.70

duc

-0.70

nect

-0.69

cast

-0.69

POSITIVE LOGITS

 sides

1.50

 sexes

1.43

 parties

1.28

 genders

1.23

 halves

1.09

 Houses

0.91

 factions

0.91

 Clintons

0.89

 spouses

0.88

 Parties

0.86

Activations Density 0.063%

pairs or groups of entities or topics discussed together

references to duality or items expressed in pairs

No Comments

No Known Activations