INDEX

Explanations

expressions of hope, assumption, imagination, prediction, and conclusion

oai_token-act-pair · gpt-3.5-turbo

phrases that express limitations on expectations or assumptions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.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

 Dill

-0.66

furt

-0.59

didn

-0.58

 Morse

-0.58

skirts

-0.57

colo

-0.57

ahead

-0.57

VERTISEMENT

-0.56

 Codes

-0.56

 methodological

-0.55

POSITIVE LOGITS

 speculate

1.04

 afford

0.99

 assume

0.95

 imagine

0.94

 hope

0.90

 pray

0.88

 succeed

0.86

 aspire

0.85

 conclude

0.84

 dream

0.82

Activations Density 0.029%

expressions of hope, assumption, imagination, prediction, and conclusion

phrases that express limitations on expectations or assumptions

No Comments

No Known Activations