INDEX

Explanations

information related to specific years or time periods

oai_token-act-pair · gpt-3.5-turbo

occurrences of the possessive form "year's."

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

eno

-0.84

loo

-0.79

der

-0.77

hari

-0.76

eller

-0.76

ucket

-0.75

then

-0.74

ther

-0.74

ova

-0.73

ghan

-0.73

POSITIVE LOGITS

 biggest

0.91

 anniversary

0.90

 inaugural

0.88

 edition

0.86

 incarnation

0.85

 greatest

0.78

 deadliest

0.76

 Greatest

0.76

 finest

0.75

 inauguration

0.73

Activations Density 0.036%

information related to specific years or time periods

occurrences of the possessive form "year's."

No Comments

No Known Activations