INDEX

Explanations

pairs of words that appear to be in the format of "verb + adverb."

oai_token-act-pair · gpt-3.5-turbo

phrases indicating relationships and connections between different elements or ideas

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

ocre

-0.73

thood

-0.67

ACY

-0.65

acies

-0.61

aram

-0.61

ngth

-0.60

ithe

-0.59

anye

-0.59

thren

-0.57

onomy

-0.56

POSITIVE LOGITS

CTR

0.72

 Clarks

0.65

etheless

0.63

 stim

0.62

 Sands

0.62

_-

0.62

 istg

0.58

 skelet

0.57

hess

0.55

 metab

0.55

Activations Density 1.063%

pairs of words that appear to be in the format of "verb + adverb."

phrases indicating relationships and connections between different elements or ideas

No Comments

No Known Activations