INDEX

Explanations

phrases related to conflicts, revenge, and drama in a narrative

oai_token-act-pair · gpt-3.5-turbo 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

chnology

-0.77

çĦ

-0.74

OPLE

-0.74

 stalls

-0.70

 Vaugh

-0.69

×Ļ×

-0.66

 sight

-0.65

IDER

-0.65

lished

-0.65

LECT

-0.65

POSITIVE LOGITS

agascar

1.16

oka

0.97

hya

0.97

onna

0.95

eline

0.95

ras

0.93

urai

0.91

rid

0.90

der

0.90

roid

0.87

Activations Density 0.014%

phrases related to conflicts, revenge, and drama in a narrative

No Comments

No Known Activations