INDEX

Explanations

ironically, the neuron appears to activate most strongly for the word "these" across a variety of contexts

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "these" 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 @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

llor

-0.72

zzle

-0.70

ppe

-0.69

pless

-0.69

achus

-0.67

-0.66

renheit

-0.66

Tank

-0.65

lest

-0.64

nah

-0.62

POSITIVE LOGITS

 fellows

0.90

 newfound

0.84

 kinds

0.77

 findings

0.77

 sorts

0.73

 guys

0.73

 proceedings

0.73

 particular

0.72

 sights

0.72

 respective

0.70

Activations Density 0.060%

ironically, the neuron appears to activate most strongly for the word "these" across a variety of contexts

occurrences of the word "these" in various contexts

No Comments

No Known Activations