INDEX

Explanations

quantitative measurements or limits

oai_token-act-pair · gpt-3.5-turbo

terms related to limits or thresholds

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

FORE

-0.79

 Guys

-0.78

ãĥīãĥ©

-0.75

bane

-0.74

hell

-0.70

ãĤ¡

-0.70

Interstitial

-0.70

alus

-0.69

alter

-0.69

by

-0.67

POSITIVE LOGITS

 allowable

1.24

 amount

1.12

 likelihood

1.02

imize

1.01

nesota

1.00

 permissible

0.96

 extent

0.95

 length

0.93

 lifespan

0.92

 allowed

0.90

Activations Density 0.024%

quantitative measurements or limits

terms related to limits or thresholds

No Comments

No Known Activations