INDEX

Explanations

words related to initiation, beginning, or starting

oai_token-act-pair · gpt-3.5-turbo

instances of the term "inst" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ikawa

-0.72

ero

-0.72

Bron

-0.68

orters

-0.68

 specificity

-0.68

rosis

-0.68

phabet

-0.67

umber

-0.67

 Rosen

-0.66

oos

-0.63

POSITIVE LOGITS

 inst

4.01

inst

1.66

Inst

1.57

 INST

1.47

 Instr

1.44

 Inst

1.38

INST

1.31

 installments

1.08

eng

1.08

 instr

1.08

Activations Density 0.011%

words related to initiation, beginning, or starting

instances of the term "inst" in various contexts

No Comments

No Known Activations