INDEX

Explanations

terms related to entry or access points

oai_token-act-pair · gpt-3.5-turbo

references to "gate" in various contexts

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

 Hots

-0.81

ortium

-0.74

TING

-0.70

arus

-0.69

ity

-0.68

lihood

-0.67

ensional

-0.66

issance

-0.65

ynasty

-0.65

ied

-0.64

POSITIVE LOGITS

keeper

1.25

keepers

1.24

ways

1.20

posts

0.98

way

0.97

fold

0.94

keeping

0.90

stones

0.89

house

0.87

hole

0.83

Activations Density 0.030%

terms related to entry or access points

references to "gate" in various contexts

No Comments

No Known Activations