INDEX

Explanations

words related to carcinogens and substances potentially harmful to health

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

gency

-0.90

ikarp

-0.68

 Haku

-0.68

ging

-0.67

mble

-0.67

guyen

-0.65

die

-0.64

 reluct

-0.63

theless

-0.63

gie

-0.63

POSITIVE LOGITS

ogens

1.02

ational

0.91

adal

0.86

eland

0.86

otta

0.86

cano

0.85

agen

0.85

ifer

0.83

olate

0.82

tern

0.81

Activations Density 0.014%

words related to carcinogens and substances potentially harmful to health

No Comments

No Known Activations