INDEX

Explanations

words related to mental health and cognitive behavior

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

reek

-0.68

git

-0.68

ICA

-0.66

 Spectre

-0.66

 Arabia

-0.65

KH

-0.64

 Mehran

-0.62

Mex

-0.61

AK

-0.61

 Antar

-0.60

POSITIVE LOGITS

 faculties

1.05

iac

0.99

fulness

0.92

storms

0.92

 disson

0.89

ostic

0.81

 disorder

0.79

ostics

0.79

stem

0.79

cube

0.78

Activations Density 2.360%

words related to mental health and cognitive behavior

No Comments

No Known Activations