INDEX

Explanations

terms related to a specific location or company called Cant

oai_token-act-pair · gpt-3.5-turbo

references to the term "cant" 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 @ 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

CVE

-0.85

 unchanged

-0.72

 prizes

-0.69

 Sharon

-0.67

ONSORED

-0.65

ATK

-0.64

CHR

-0.63

heads

-0.63

 Sandy

-0.61

ç·

-0.61

POSITIVE LOGITS

rell

1.34

een

1.27

illon

1.25

inas

1.10

onal

1.08

illation

1.06

ina

1.04

ones

1.00

icle

0.99

rip

0.98

Activations Density 0.028%

terms related to a specific location or company called Cant

references to the term "cant" in various contexts

No Comments

No Known Activations