INDEX

Explanations

phrases related to public relations (PR) activities

oai_token-act-pair · gpt-3.5-turbo

references to a specific program or initiative identified by "PR."

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

lihood

-0.91

 Resurrection

-0.77

 Genesis

-0.76

 Presbyter

-0.73

mates

-0.72

 Tale

-0.72

lings

-0.71

 Dwarf

-0.68

querque

-0.67

ãĥĥãĤ¯

-0.65

POSITIVE LOGITS

INC

1.16

ODUCT

1.12

IME

1.08

IZE

1.05

ISM

1.02

INTER

1.01

OPER

0.99

INT

0.99

IOR

0.98

OF

0.89

Activations Density 0.011%

phrases related to public relations (PR) activities

references to a specific program or initiative identified by "PR."

No Comments

No Known Activations