INDEX

Explanations

references to numbers and lists

oai_token-act-pair · gpt-3.5-turbo

numerical rankings or order in lists

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

 behavi

-0.81

 authority

-0.76

 tremend

-0.74

 behav

-0.73

 inward

-0.67

uggest

-0.67

izons

-0.65

 moder

-0.65

 rule

-0.65

 heroine

-0.64

POSITIVE LOGITS

 Learns

0.81

Bye

0.78

 Provided

0.78

TBD

0.77

 Scores

0.75

 Unknown

0.71

YES

0.70

0.69

adem

0.69

Activations Density 0.117%

references to numbers and lists

numerical rankings or order in lists

No Comments

No Known Activations