INDEX

Explanations

words related to a specific collection or group of items

oai_token-act-pair · gpt-3.5-turbo

references to groups or collections of items or concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

IBLE

-0.74

Fight

-0.72

 Souls

-0.71

 Authority

-0.70

æĺ¯

-0.69

Sense

-0.68

 sincerity

-0.65

 charism

-0.64

ï¸ı

-0.64

 Biblical

-0.63

POSITIVE LOGITS

tle

0.99

offs

0.96

table

0.88

list

0.86

atum

0.85

meal

0.84

estone

0.83

chell

0.78

 forth

0.76

itia

0.75

Activations Density 0.033%

words related to a specific collection or group of items

references to groups or collections of items or concepts

No Comments

No Known Activations