INDEX

Explanations

terms related to surpassing or exceeding something

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "overt" and its variations

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

nesota

-0.77

 Crate

-0.72

UTION

-0.72

 Caps

-0.70

lihood

-0.67

 Trident

-0.63

 Bucket

-0.63

 Borough

-0.62

 Buffer

-0.61

 Union

-0.61

POSITIVE LOGITS

 overt

0.98

imes

0.97

uned

0.95

aken

0.89

empt

0.81

itled

0.81

aking

0.80

orical

0.79

akes

0.76

orial

0.76

Activations Density 0.007%

terms related to surpassing or exceeding something

occurrences of the word "overt" and its variations

No Comments

No Known Activations