INDEX

Explanations

mentions of being selected for something, whether it be a position, program, or opportunity

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "selected" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

cre

-0.79

 heels

-0.75

pl

-0.73

 sink

-0.72

forth

-0.68

LI

-0.67

per

-0.66

cing

-0.66

Elsa

-0.65

ced

-0.65

POSITIVE LOGITS

 selected

1.09

selected

1.00

 selections

0.97

 selects

0.94

selection

0.93

 selection

0.90

 Selection

0.89

avorite

0.88

 select

0.85

ecided

0.84

Activations Density 0.010%

mentions of being selected for something, whether it be a position, program, or opportunity

occurrences of the word "selected" in various contexts

No Comments

No Known Activations