INDEX

Explanations

mathematical terms related to calculations

oai_token-act-pair · gpt-3.5-turbo

references to the word "ore."

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

soDeliveryDate

-0.68

srf

-0.68

¥ŀ

-0.63

ulk

-0.63

 truce

-0.62

uates

-0.61

 slightest

-0.61

 privileged

-0.60

 disemb

-0.59

channelAvailability

-0.59

POSITIVE LOGITS

tto

1.07

gon

0.98

lli

0.94

llo

0.91

tsky

0.91

lette

0.90

lett

0.90

tta

0.89

ore

0.83

nz

0.82

Activations Density 0.007%

mathematical terms related to calculations

references to the word "ore."

No Comments

No Known Activations