INDEX

Explanations

phrases related to self-improvement and accountability

oai_token-act-pair · gpt-3.5-turbo

themes related to positivity and encouragement

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

 bona

-0.79

 godd

-0.78

 Hemisphere

-0.77

 goddamn

-0.75

 egreg

-0.74

 Godd

-0.74

 promptly

-0.73

 subpoena

-0.72

 authorize

-0.71

 subpoen

-0.70

POSITIVE LOGITS

 loneliness

1.19

 counselling

1.11

 youths

1.10

 behaviours

1.09

 behavioural

1.05

 emphas

1.05

oneliness

1.00

 pupils

0.98

 programmes

0.98

 migrants

0.97

Activations Density 1.059%

phrases related to self-improvement and accountability

themes related to positivity and encouragement

No Comments

No Known Activations