INDEX

Explanations

references to specific episodes from a series or show

oai_token-act-pair · gpt-3.5-turbo

references to specific episodes of television shows

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

minist

-0.91

folk

-0.80

 denomin

-0.75

sheets

-0.74

ndra

-0.71

uckland

-0.70

ensical

-0.69

gie

-0.68

press

-0.68

none

-0.67

POSITIVE LOGITS

 Episode

1.04

 Transcript

0.83

VII

0.77

Episode

0.77

 Tickets

0.72

ÃįÃį

0.72

 Creator

0.72

 Flavoring

0.71

Stage

0.70

 Course

0.70

Activations Density 0.026%

references to specific episodes from a series or show

references to specific episodes of television shows

No Comments

No Known Activations