INDEX

Explanations

phrases related to recurring patterns or themes

oai_token-act-pair · gpt-3.5-turbo

recurring themes and patterns in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

poon

-0.83

ateurs

-0.79

 ceremon

-0.75

 sacrific

-0.73

atography

-0.73

apest

-0.69

aband

-0.68

 pets

-0.67

unte

-0.66

cemic

-0.66

POSITIVE LOGITS

 underlying

0.96

 endemic

0.89

 occurring

0.87

 pattern

0.87

 plag

0.86

 generational

0.85

 phenomenon

0.85

 metast

0.85

 psychosis

0.85

 phenomena

0.83

Activations Density 0.488%

phrases related to recurring patterns or themes

recurring themes and patterns in various contexts

No Comments

No Known Activations