INDEX

Explanations

mentions of the word "studio"

oai_token-act-pair · gpt-3.5-turbo

references to a specific studio

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

abad

-0.74

pend

-0.71

idem

-0.66

MpServer

-0.66

andum

-0.64

avez

-0.62

ocaust

-0.62

mere

-0.62

bh

-0.61

theless

-0.60

POSITIVE LOGITS

 studio

0.96

 studios

0.90

mates

0.79

orpor

0.77

yrinth

0.72

floor

0.71

 apartment

0.71

 floor

0.71

 mates

0.70

ounter

0.70

Activations Density 0.042%

mentions of the word "studio"

references to a specific studio

No Comments

No Known Activations