INDEX

Explanations

phrases related to consequences, actions, and decision-making

oai_token-act-pair · gpt-3.5-turbo

expressions of caution or concern about actions and their consequences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

htaking

-0.69

interstitial

-0.67

orious

-0.61

 Built

-0.60

 Skin

-0.58

æľ

-0.56

 fame

-0.56

itled

-0.55

culosis

-0.54

 Birth

-0.53

POSITIVE LOGITS

 defe

0.70

quo

0.66

 anyway

0.64

etheless

0.64

 answ

0.64

 administr

0.62

 uncertainties

0.61

 downstream

0.61

 redund

0.60

 inaction

0.59

Activations Density 2.781%

phrases related to consequences, actions, and decision-making

expressions of caution or concern about actions and their consequences

No Comments

No Known Activations