INDEX

Explanations

mentions of specific tasks or actions related to handling responsibilities

oai_token-act-pair · gpt-3.5-turbo

phrases indicating strong reactions or outcomes related to accidents and penalties

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 hampered

-0.60

ource

-0.58

 inconsist

-0.56

catentry

-0.56

 injuries

-0.55

 aliases

-0.55

 latter

-0.54

 enhancements

-0.54

 severity

-0.53

 indirect

-0.53

POSITIVE LOGITS

 someday

0.86

 tomorrow

0.77

ä»£

0.71

insert

0.69

 whichever

0.66

AAAAAAAA

0.66

 anytime

0.66

ãĥĺ

0.62

bucks

0.60

Imagine

0.58

Activations Density 1.577%

mentions of specific tasks or actions related to handling responsibilities

phrases indicating strong reactions or outcomes related to accidents and penalties

No Comments

No Known Activations