INDEX

Explanations

phrases related to providing evidence or support for statements

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate the need for or presence of support and validation

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

_>

-0.86

erves

-0.73

====

-0.72

umn

-0.70

mouth

-0.70

phas

-0.67

âĺħ

-0.67

bell

-0.66

Sport

-0.65

jee

-0.65

POSITIVE LOGITS

IFT

0.70

rison

0.69

river

0.63

dates

0.62

plane

0.60

stairs

0.59

stead

0.59

lishes

0.59

stroke

0.59

 assaults

0.58

Activations Density 0.016%

phrases related to providing evidence or support for statements

phrases that indicate the need for or presence of support and validation

No Comments

No Known Activations