INDEX

Explanations

phrases related to seriousness or taking something seriously

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

tein

-0.83

inia

-0.82

artifacts

-0.72

lyn

-0.72

anic

-0.69

orius

-0.69

ramid

-0.69

rious

-0.67

keys

-0.66

irlwind

-0.65

POSITIVE LOGITS

 enough

0.95

 dent

0.73

 seriously

0.72

 seriousness

0.69

terday

0.66

 journalistic

0.64

ATIONAL

0.61

NESS

0.60

 advis

0.60

 internally

0.60

Activations Density 0.017%

phrases related to seriousness or taking something seriously

No Comments

No Known Activations