INDEX

Explanations

words related to mental health issues, medication, and online markets

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

ingen

-0.77

vous

-0.71

ned

-0.70

ified

-0.68

ding

-0.66

soDeliveryDate

-0.66

 Kardashian

-0.63

risome

-0.63

eners

-0.63

bered

-0.63

POSITIVE LOGITS

illary

1.07

alon

0.90

qua

0.85

osta

0.85

oust

0.84

iom

0.84

venture

0.81

ansas

0.81

ibaba

0.80

ħĭ

0.79

Activations Density 3.215%

words related to mental health issues, medication, and online markets

No Comments

No Known Activations