INDEX

Explanations

phrases indicating confusion or a need for understanding

oai_token-act-pair · gpt-3.5-turbo

references to the question "why."

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

lator

-0.80

amps

-0.72

 Roller

-0.72

ymph

-0.71

ãĤ¤ãĥĪ

-0.71

aughed

-0.68

phrine

-0.66

rop

-0.63

aire

-0.62

luck

-0.61

POSITIVE LOGITS

why

0.98

WHY

0.89

soever

0.86

why

0.85

 exactly

0.80

Why

0.69

Origin

0.67

 justifying

0.63

 bother

0.62

iterranean

0.62

Activations Density 0.032%

phrases indicating confusion or a need for understanding

references to the question "why."

No Comments

No Known Activations