INDEX

Explanations

words related to ongoing actions or processes

oai_token-act-pair · gpt-3.5-turbo

words that describe states of being or actions typically associated with a process of queuing or delays

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

osponsors

-0.68

TPP

-0.68

units

-0.64

roma

-0.63

tein

-0.62

flies

-0.61

cy

-0.61

OD

-0.61

 trenches

-0.59

bott

-0.59

POSITIVE LOGITS

ued

1.31

uing

1.12

ues

0.95

eer

0.91

uez

0.87

mble

0.84

itionally

0.83

roit

0.80

aily

0.80

vre

0.80

Activations Density 0.005%

words related to ongoing actions or processes

words that describe states of being or actions typically associated with a process of queuing or delays

No Comments

No Known Activations