INDEX

Explanations

negative statements or warnings about actions or situations

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

negations or phrases that imply doubt or uncertainty

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

 millenn

-0.94

 adolesc

-0.93

senal

-0.93

¿½

-0.92

PDATE

-0.88

 Palestin

-0.87

WithNo

-0.86

ò

-0.83

Fra

-0.81

ñ

-0.80

POSITIVE LOGITS

 necessarily

1.32

be

1.16

 realize

1.07

yet

1.03

 exactly

1.03

 always

1.03

 even

1.00

 entirely

0.96

 quite

0.96

 seem

0.95

Activations Density 0.071%

negative statements or warnings about actions or situations

negations or phrases that imply doubt or uncertainty

No Comments

No Known Activations