INDEX

Explanations

phrases related to investigations or scandals

oai_token-act-pair · gpt-3.5-turbo

rhetorical questions

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

 migration

-0.74

 glyph

-0.70

 migr

-0.68

 targets

-0.65

 Manila

-0.64

atche

-0.64

 inactive

-0.63

 analges

-0.63

 wagon

-0.61

 listens

-0.61

POSITIVE LOGITS

Answer

1.41

?????-?????-

1.03

ãĥ¯ãĥ³

1.01

è¦ļéĨĴ

0.95

 Answer

0.92

ãĤ¢ãĥ«

0.91

?????-

0.90

ãĥ¤

0.89

inventoryQuantity

0.85

?????

0.85

Activations Density 0.232%

phrases related to investigations or scandals

rhetorical questions

No Comments

No Known Activations