INDEX

Explanations

instances discussing reaching or achieving specific levels, targets, or goals in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to achieving or reaching goals and potential

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

antine

-0.72

behind

-0.71

swer

-0.71

ocument

-0.71

herty

-0.69

vertisements

-0.65

struct

-0.65

 Corpus

-0.64

ry

-0.62

gov

-0.62

POSITIVE LOGITS

 fullest

1.15

 rightful

1.03

 stride

1.02

own

0.96

 allotted

0.94

 destination

0.93

 destiny

0.91

 greatest

0.89

 quota

0.88

 limits

0.88

Activations Density 0.085%

instances discussing reaching or achieving specific levels, targets, or goals in various contexts

references to achieving or reaching goals and potential

No Comments

No Known Activations