INDEX

Explanations

words related to ultimate outcomes or conclusions

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

 Recently

-0.66

Wee

-0.64

kh

-0.64

RN

-0.63

Normally

-0.62

 Frog

-0.62

 Flying

-0.62

 formerly

-0.61

maid

-0.60

 Yesterday

-0.59

POSITIVE LOGITS

 deciding

0.81

 decide

0.81

 culmin

0.79

 boils

0.77

 decided

0.76

 unsuccessful

0.75

 succeed

0.73

 succumb

0.72

apor

0.72

urable

0.71

Activations Density 0.024%

words related to ultimate outcomes or conclusions

No Comments

No Known Activations