INDEX

Explanations

names and titles associated with crimes or paranormal activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

ADRA

-0.70

Tokens

-0.66

ÙĴ

-0.65

âĸ¬âĸ¬

-0.65

Story

-0.64

WATCH

-0.63

FIX

-0.63

 overwhelming

-0.62

 asses

-0.62

 manufacturing

-0.61

POSITIVE LOGITS

oglu

1.13

opoulos

1.02

ski

1.02

sson

0.98

ewski

0.97

zyk

0.97

icz

0.94

oÄŁ

0.92

sky

0.92

oulos

0.90

Activations Density 0.114%

names and titles associated with crimes or paranormal activities

No Comments

No Known Activations