INDEX

Explanations

words related to things or concepts that are external, separate, or different from a particular context

oai_token-act-pair · gpt-3.5-turbo

references to the concept of being "outside" or external to a particular context

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

cca

-0.78

chief

-0.76

enegger

-0.76

lda

-0.72

killer

-0.72

ulous

-0.70

thritis

-0.68

anche

-0.67

brance

-0.67

uncle

-0.66

POSITIVE LOGITS

shock

0.82

etheless

0.80

 observer

0.75

 Nicaragua

0.71

outside

0.71

 observers

0.70

side

0.68

 academia

0.67

 Indies

0.67

ileged

0.67

Activations Density 0.024%

words related to things or concepts that are external, separate, or different from a particular context

references to the concept of being "outside" or external to a particular context

No Comments

No Known Activations