INDEX

Explanations

phrases related to reasons or causes

oai_token-act-pair · gpt-3.5-turbo

pronouns or references indicating possession and existence in statements or contexts

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

wikipedia

-0.82

jri

-0.70

ption

-0.69

thood

-0.68

acs

-0.66

itudes

-0.66

Anne

-0.66

Ni

-0.65

ocation

-0.64

zac

-0.63

POSITIVE LOGITS

 lacked

0.85

 lacks

0.82

 lack

0.80

 didn

0.79

 cumbers

0.78

 hadn

0.76

 messed

0.74

 weren

0.71

 dared

0.70

 knew

0.70

Activations Density 0.205%

phrases related to reasons or causes

pronouns or references indicating possession and existence in statements or contexts

No Comments

No Known Activations