INDEX

Explanations

phrases indicating addition or accumulation

oai_token-act-pair · gpt-3.5-turbo

phrases that accumulate or build upon previous information or context

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

glers

-0.80

ptin

-0.72

heres

-0.69

cock

-0.68

araoh

-0.68

borg

-0.67

ata

-0.67

RESULTS

-0.65

 ______

-0.65

hots

-0.64

POSITIVE LOGITS

 equation

0.94

 coffers

0.88

 detriment

0.86

 list

0.86

 ranks

0.83

 realm

0.81

 existing

0.79

 repertoire

0.78

 ende

0.78

ortment

0.77

Activations Density 0.210%

phrases indicating addition or accumulation

phrases that accumulate or build upon previous information or context

No Comments

No Known Activations