INDEX

Explanations

references to the telecommunications company AT&T

oai_token-act-pair · gpt-3.5-turbo

occurrences of the abbreviation "AT&T."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 pupils

-0.69

erva

-0.67

 vomiting

-0.67

 academ

-0.66

 diam

-0.62

 Croatian

-0.62

 Libyan

-0.61

 Haitian

-0.61

 prose

-0.60

 liquid

-0.60

POSITIVE LOGITS

ENA

0.86

HI

0.82

Wan

0.79

REE

0.79

Mobile

0.77

rex

0.77

IB

0.76

tin

0.76

ACY

0.75

NI

0.74

Activations Density 0.029%

references to the telecommunications company AT&T

occurrences of the abbreviation "AT&T."

No Comments

No Known Activations