INDEX

Explanations

emphatic responses or confirmations

oai_token-act-pair · gpt-3.5-turbo

punctuation marks, especially colons

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

 unification

-0.73

 rule

-0.70

 caster

-0.69

 reckoning

-0.67

 downstream

-0.66

 consequential

-0.64

 upstream

-0.64

avorite

-0.63

 baskets

-0.62

 announ

-0.61

POSITIVE LOGITS

 Yeah

1.03

Bye

0.90

 Alright

0.86

 Absolutely

0.82

Yeah

0.81

 Okay

0.80

Huh

0.78

Oh

0.78

Hmm

0.77

Yep

0.75

Activations Density 0.080%

emphatic responses or confirmations

punctuation marks, especially colons

No Comments

No Known Activations