INDEX

Explanations

the word "chosen" with a high level of activation

oai_token-act-pair · gpt-3.5-turbo

instances of the word "chosen."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

pat

-0.76

Net

-0.75

urst

-0.71

ptoms

-0.70

ilit

-0.67

vacc

-0.67

CDC

-0.66

monds

-0.66

itamin

-0.65

emic

-0.65

POSITIVE LOGITS

 chosen

1.01

 randomly

0.83

 chooses

0.80

lists

0.80

 choosing

0.78

 chose

0.75

 ACTIONS

0.75

 Disciple

0.74

çĶŁ

0.73

selection

0.71

Activations Density 0.010%

the word "chosen" with a high level of activation

instances of the word "chosen."

No Comments

No Known Activations