INDEX

Explanations

software-related terms related to fixing bugs or issues

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

SPONSORED

-1.04

é¾įåĸļå£«

-0.81

Redditor

-0.79

ãĤ¼ãĤ¦ãĤ¹

-0.76

ãĥĺ

-0.74

arious

-0.74

hander

-0.73

tnc

-0.72

mop

-0.72

writers

-0.72

POSITIVE LOGITS

 typo

0.81

 Fixes

0.80

 erroneous

0.75

 reversible

0.74

 wording

0.71

 procedural

0.67

 cler

0.67

UI

0.66

 errone

0.66

bug

0.66

Activations Density 0.081%

software-related terms related to fixing bugs or issues

No Comments

No Known Activations