INDEX

Explanations

phrases where the word "how" appears

oai_token-act-pair · gpt-3.5-turbo

questions and phrases indicating uncertainty or inquiries about methods

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

 resil

-0.68

emetery

-0.64

 cardinal

-0.64

 srfAttach

-0.64

 lock

-0.61

 Pirates

-0.61

Ley

-0.61

 commencement

-0.61

 Cron

-0.61

 redundancy

-0.61

POSITIVE LOGITS

theless

0.97

hee

0.89

soever

0.80

here

0.78

ean

0.78

tenance

0.77

selves

0.75

fter

0.74

cation

0.74

ians

0.73

Activations Density 0.031%

phrases where the word "how" appears

questions and phrases indicating uncertainty or inquiries about methods

No Comments

No Known Activations