INDEX

Explanations

dates specified as 'Friday'

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "Friday."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 prolifer

-0.66

hess

-0.65

 veter

-0.65

arth

-0.64

 prototyp

-0.62

lik

-0.61

glomer

-0.61

otin

-0.61

hesis

-0.61

erest

-0.60

POSITIVE LOGITS

 Friday

3.32

 Thursday

2.85

Friday

2.80

 Saturday

2.72

 Wednesday

2.59

 Monday

2.53

 Tuesday

2.53

 Sunday

2.46

Thursday

2.38

 Fridays

2.27

Activations Density 0.022%

dates specified as 'Friday'

occurrences of the word "Friday."

No Comments

No Known Activations