INDEX

Explanations

references to scholarly studies or reports

oai_token-act-pair · gpt-3.5-turbo

references to reports, studies, or proposals, often accompanied by numerical data or important descriptors

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

unic

-0.69

pected

-0.66

ateurs

-0.65

peg

-0.63

Switch

-0.63

isp

-0.62

ge

-0.62

pec

-0.61

iple

-0.61

igmatic

-0.60

POSITIVE LOGITS

 however

1.04

 namely

0.94

 albeit

0.91

 wherein

0.90

 which

0.90

 according

0.84

 citing

0.83

 comprising

0.82

 whereby

0.79

 titled

0.78

Activations Density 0.413%

references to scholarly studies or reports

references to reports, studies, or proposals, often accompanied by numerical data or important descriptors

No Comments

No Known Activations