INDEX

Explanations

specific keywords indicating a temporal sequence or context transition

oai_token-act-pair · gpt-3.5-turbo

the word "Before" indicating prior events 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

aren

-0.83

utter

-0.79

pez

-0.68

wire

-0.66

ILY

-0.65

è¦ļéĨĴ

-0.65

hyde

-0.65

hack

-0.63

erry

-0.63

amount

-0.62

POSITIVE LOGITS

cluding

0.75

rely

0.72

pping

0.72

irement

0.71

 realizing

0.71

 Stats

0.70

noon

0.70

 anyone

0.66

 concluding

0.64

hand

0.64

Activations Density 0.035%

specific keywords indicating a temporal sequence or context transition

the word "Before" indicating prior events or contexts

No Comments

No Known Activations