INDEX

Explanations

words related to danger or threats in a narrative context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ackets

-0.81

 Zucker

-0.78

Article

-0.76

insky

-0.73

 MSNBC

-0.73

Newsletter

-0.72

orters

-0.71

Romney

-0.71

orter

-0.70

Sports

-0.70

POSITIVE LOGITS

 sorcerer

1.06

 sorce

1.06

 treacher

1.05

 corrupted

1.04

 undead

1.04

 demonic

1.02

 cursed

1.01

 Norn

1.01

 Aether

0.99

 prophes

0.98

Activations Density 0.804%

words related to danger or threats in a narrative context

No Comments

No Known Activations