INDEX

Explanations

words related to problem-solving, decision-making, and challenges

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

nesday

-0.78

cale

-0.77

yip

-0.69

ynski

-0.68

hirt

-0.67

rower

-0.67

uden

-0.65

ndum

-0.63

ync

-0.62

uckland

-0.62

POSITIVE LOGITS

breakers

0.90

 makers

0.89

makers

0.88

 sizes

0.87

able

0.82

ings

0.82

 cancell

0.81

 theoret

0.81

holders

0.80

boards

0.80

Activations Density 0.578%

words related to problem-solving, decision-making, and challenges

No Comments

No Known Activations