INDEX

Explanations

words related to the effectiveness or efficacy of treatments and medical practices

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

pper

-0.71

ODE

-0.70

 Wonderland

-0.68

Pic

-0.67

dit

-0.65

houn

-0.65

 Federation

-0.62

 Institution

-0.62

hak

-0.61

 Brotherhood

-0.60

POSITIVE LOGITS

iveness

1.20

iencies

1.06

 effectiveness

1.00

fulness

0.96

acies

0.94

acy

0.92

 tremend

0.91

abilities

0.90

ality

0.85

ulence

0.85

Activations Density 0.011%

words related to the effectiveness or efficacy of treatments and medical practices

No Comments

No Known Activations