INDEX

Explanations

actions related to customer interaction and user experience

oai_token-act-pair · gpt-3.5-turbo

phrases indicating user engagement and accessibility

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

catentry

-0.74

 predecessor

-0.64

Cooldown

-0.60

 successor

-0.59

ãĥ¯

-0.57

ãĤ´ãĥ³

-0.55

rul

-0.55

 counterpart

-0.53

 coincidence

-0.52

Translation

-0.52

POSITIVE LOGITS

 themselves

0.79

 their

0.64

 THEIR

0.63

 elect

0.61

 whom

0.60

selves

0.60

 voluntarily

0.59

 alike

0.58

their

0.56

 revolt

0.56

Activations Density 1.203%

actions related to customer interaction and user experience

phrases indicating user engagement and accessibility

No Comments

No Known Activations