INDEX

Explanations

negative statements indicating an outcome that is contrary to expectation

oai_token-act-pair · gpt-3.5-turbo

negative qualifiers and phrases indicating uncertainty or negation

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

assembly

-0.68

 Rebellion

-0.66

IAS

-0.64

mining

-0.61

¿½

-0.61

WithNo

-0.60

holding

-0.59

çĶŁ

-0.58

 Registry

-0.58

LET

-0.58

POSITIVE LOGITS

 necessarily

1.15

be

1.05

 realize

0.94

 agree

0.89

necess

0.88

 realise

0.87

 have

0.85

 qualify

0.84

 seem

0.83

 even

0.82

Activations Density 0.062%

negative statements indicating an outcome that is contrary to expectation

negative qualifiers and phrases indicating uncertainty or negation

No Comments

No Known Activations