INDEX

Explanations

width measurements in pixels or percentages

oai_token-act-pair · gpt-3.5-turbo

specifications related to dimensions and measurements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

andum

-0.86

plet

-0.83

Ã³n

-0.72

ruption

-0.70

onto

-0.70

inances

-0.70

ogical

-0.67

dinand

-0.67

ovie

-0.67

ectomy

-0.66

POSITIVE LOGITS

sheets

0.91

sheet

0.90

ï¸

0.85

0.84

span

0.83

sd

0.82

sw

0.81

 width

0.81

pread

0.80

 Width

0.79

Activations Density 0.022%

width measurements in pixels or percentages

specifications related to dimensions and measurements

No Comments

No Known Activations