INDEX

Explanations

the word "consider."

oai_token-act-pair · gpt-3.5-turbo

directives or prompts to reflect on actions or decisions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

Els

-0.76

phys

-0.74

aqu

-0.71

vous

-0.70

zona

-0.70

oiler

-0.69

inos

-0.67

WARN

-0.67

PATH

-0.66

IER

-0.64

POSITIVE LOGITS

 whether

1.10

 donating

1.03

 abandoning

0.94

 oneself

0.90

 aloud

0.89

 suing

0.88

 joining

0.88

 alternatives

0.87

 withdrawing

0.86

 yourselves

0.86

Activations Density 0.044%

the word "consider."

directives or prompts to reflect on actions or decisions

No Comments

No Known Activations