INDEX

Explanations

statements or beliefs that apply universally to a group or situation, often involving a contrast or comparison

oai_token-act-pair · gpt-3.5-turbo

phrases indicating disagreement or differing opinions

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

uggest

-0.75

 newcom

-0.70

 Hover

-0.69

 ................

-0.65

 instead

-0.64

 actionGroup

-0.64

 lest

-0.63

 rather

-0.62

none

-0.62

unknown

-0.60

POSITIVE LOGITS

 necessarily

0.97

 agree

0.90

 agrees

0.86

baum

0.81

ple

0.74

 fortunate

0.73

 reacts

0.73

 subscrib

0.71

estone

0.71

 equally

0.70

Activations Density 0.147%

statements or beliefs that apply universally to a group or situation, often involving a contrast or comparison

phrases indicating disagreement or differing opinions

No Comments

No Known Activations