INDEX

Explanations

phrases related to completion or finishing a task

oai_token-act-pair · gpt-3.5-turbo

phrases indicating completion or tasks that have been finished

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

thur

-0.78

lights

-0.78

ussen

-0.77

wit

-0.74

liner

-0.73

Reviewer

-0.73

bug

-0.70

rum

-0.68

dayName

-0.68

lua

-0.67

POSITIVE LOGITS

pez

1.07

 toget

0.79

 done

0.76

escription

0.71

omsday

0.71

omething

0.69

 mosqu

0.68

 differently

0.67

 outreach

0.67

Īè

0.67

Activations Density 0.027%

phrases related to completion or finishing a task

phrases indicating completion or tasks that have been finished

No Comments

No Known Activations