INDEX

Explanations

expressions or mentions of gratitude towards support

oai_token-act-pair · gpt-3.5-turbo

expressions of gratitude and mentions of support

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ãĥ£

-0.70

kered

-0.66

vern

-0.63

 pores

-0.63

iren

-0.63

 Hebdo

-0.63

 sweat

-0.62

 contrad

-0.61

 dare

-0.60

 unbeliev

-0.60

POSITIVE LOGITS

Support

0.78

 support

0.77

hesis

0.75

heses

0.75

orship

0.75

bands

0.75

 Supports

0.75

enza

0.74

asio

0.72

ament

0.71

Activations Density 0.051%

expressions or mentions of gratitude towards support

expressions of gratitude and mentions of support

No Comments

No Known Activations