INDEX

Explanations

words related to physical features or actions

oai_token-act-pair · gpt-3.5-turbo

variations of the letters associated with specific repetitive sounds or patterns

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

 Dangerous

-0.70

 Human

-0.70

è£ħ

-0.65

Nu

-0.65

 Rabbit

-0.64

 Verse

-0.64

Cla

-0.64

 Deity

-0.64

 Principle

-0.64

Pwr

-0.64

POSITIVE LOGITS

alon

1.04

utters

1.04

airs

1.02

atching

1.00

ards

0.99

ords

0.98

aults

0.98

akery

0.98

alling

0.96

urd

0.95

Activations Density 0.120%

words related to physical features or actions

variations of the letters associated with specific repetitive sounds or patterns

No Comments

No Known Activations