INDEX

Explanations

contractions common in English such as "don't" or "didn't"

oai_token-act-pair · gpt-3.5-turbo

negations related to necessity or obligation

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Adin

-0.67

forms

-0.66

 integrity

-0.63

Intern

-0.60

 Integrity

-0.59

 Strikes

-0.57

 affinity

-0.57

heartedly

-0.57

 Alas

-0.57

Fu

-0.55

POSITIVE LOGITS

 need

1.07

 hear

0.96

 gotta

0.92

 wanna

0.90

 NEED

0.88

 expect

0.85

 yourselves

0.83

 want

0.83

 yourself

0.81

 necessarily

0.81

Activations Density 0.123%

contractions common in English such as "don't" or "didn't"

negations related to necessity or obligation

No Comments

No Known Activations