INDEX

Explanations

dates in a specific format

oai_token-act-pair · gpt-3.5-turbo

negative numerical values or trends

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 moder

-0.74

 fright

-0.71

 hypers

-0.70

 slowing

-0.69

 churn

-0.68

 skim

-0.68

 human

-0.67

hon

-0.67

 executive

-0.67

 pipeline

-0.66

POSITIVE LOGITS

cv

1.26

_-

1.18

?????-

1.16

Apr

1.14

1.13

1.11

generic

1.10

sama

1.08

years

1.07

1.05

Activations Density 0.062%

dates in a specific format

negative numerical values or trends

No Comments

No Known Activations