INDEX

Explanations

words related to gaining understanding or knowledge

oai_token-act-pair · gpt-3.5-turbo

references to gaining or providing insight on various topics

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

ategory

-0.69

shr

-0.64

 fuss

-0.63

 disbanded

-0.60

 notor

-0.60

ony

-0.60

 slapping

-0.59

senal

-0.59

FACE

-0.56

 stiff

-0.56

POSITIVE LOGITS

ibility

1.00

ively

0.89

 glean

0.89

 insight

0.87

fully

0.83

 insights

0.82

ibly

0.82

spection

0.82

ible

0.81

ives

0.79

Activations Density 0.027%

words related to gaining understanding or knowledge

references to gaining or providing insight on various topics

No Comments

No Known Activations