INDEX

Explanations

mentions of the word "iter" at a high activation level

oai_token-act-pair · gpt-3.5-turbo

instances of the term "iter" 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 @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ership

-0.91

¥µ

-0.75

ecause

-0.75

itionally

-0.72

¬¼

-0.71

ruciating

-0.70

sterdam

-0.70

cipled

-0.69

orld

-0.68

yssey

-0.67

POSITIVE LOGITS

ariat

0.93

ocene

0.85

iter

0.81

imation

0.78

imate

0.78

aton

0.77

ror

0.75

asury

0.75

ATURE

0.75

ason

0.74

Activations Density 0.023%

mentions of the word "iter" at a high activation level

instances of the term "iter" in various contexts

No Comments

No Known Activations