INDEX

Explanations

references to a specific location, potentially related to culture or language

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "Cant" in various contexts, presumably related to a specific cultural or linguistic reference

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Reviewer

-0.81

ãģį

-0.71

edged

-0.66

CVE

-0.66

 mosqu

-0.64

iza

-0.64

heads

-0.64

vernment

-0.64

*/(

-0.63

æł

-0.63

POSITIVE LOGITS

illon

1.02

 Cant

0.96

onso

0.96

rell

0.87

een

0.86

aleb

0.83

esian

0.78

rip

0.78

ina

0.78

omp

0.76

Activations Density 0.007%

references to a specific location, potentially related to culture or language

mentions of the word "Cant" in various contexts, presumably related to a specific cultural or linguistic reference

No Comments

No Known Activations