INDEX

Explanations

words related to safety, fairness, and collaboration

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

chet

-0.76

berries

-0.71

yi

-0.71

planes

-0.71

chu

-0.70

 painfully

-0.68

pain

-0.68

 furiously

-0.67

yss

-0.66

stals

-0.66

POSITIVE LOGITS

tarian

0.93

GoldMagikarp

0.84

ably

0.80

acies

0.77

enough

0.76

itably

0.72

 isot

0.71

inventoryQuantity

0.70

ufact

0.69

\\\\\\\\

0.69

Activations Density 2.220%

words related to safety, fairness, and collaboration

No Comments

No Known Activations